Sev met a guy named Phillip on the Downtown Eastside. Bloody nose, black eye, the pavement behind him drenched in blood. He asked what happened. Phillip just…
Notes & highlights
Published: · 16:13
Sev met a guy named Phillip on the Downtown Eastside. Bloody nose, black eye, the pavement behind him drenched in blood. He asked what happened. Phillip just said "I got beaten up." Matter of fact. Like that's just how life is.
They started talking about computer viruses. Process trees. Virus signatures. One of the most technical conversations Sev had all year — more technical than most of his conversations at tech conferences. Phillip was a Canada Post worker before he lost his job and ended up on the street.
That encounter cracked something open.
Sev realized existing ethics treats people like isolated agents — like what you do inside your own head is all that matters. But we're nodes in a network. What flows through you — what you absorb, what you pass on — that's the actual moral work.
He built a framework called Transmutarianism. There's a formula: Moral Work = Σ(filtering + amplification) × cultural coefficient. It applies to humans, organizations, and AI agents. You can calculate whether an AI system does net positive or net negative moral work.
Then he took it to the streets. 51 interviews with Vancouver's unhoused population. The result? 63% said they want AI to show compassion, care, love, empathy. Compare that to Anthropic's 81,000-person survey of knowledge workers — they wanted productivity. Completely different priorities. Those voices are absent from training data.
————
💥 Join the movement
Join BC + AI: https://bc-ai.ca/membership/
Next Vancouver AI meetup: https://luma.com/vancouver-ai
Smash that Subscribe & 🔔 so you don't miss future talks.
Drop a comment: What's one question you'd ask someone whose voice isn't in AI training data?
Share the vid with someone working on AI ethics.